[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: a small patch
[Thread Prev] | [Thread Next]
- Subject: Re: a small patch
- From: Aris Adamantiadis <aris@xxxxxxxxxxxx>
- Reply-to: libssh@xxxxxxxxxx
- Date: Wed, 21 Apr 2010 11:08:09 +0200
- To: libssh@xxxxxxxxxx
Hi Xi, Thanks for your patch. Libssh master is currently in architectural rework and SSH1 has been a little less polished than the rest. I'm happy you managed to make it work. I have yet to verify that the \r\n -> \n thing does not break anything on openssh and cisco SSH. Regards, Aris Xi Wang a écrit : > Cool. Here goes a few more fixes when I was trying libssh with > guest@xxxxxxxxxxx, an ssh-based forum using v1. > > * banner/identification: ssh_send_banner (client.c). libssh reports > crc error after sending the session key. The actual problem is that > libssh uses "...\r\n" in its identification, while some sever doesn't > parse the identification correctly and stops at "\r". Since the ssh > specification (http://www.snailbook.com/docs/protocol-1.5.txt) says it > should be "...\n", the patch removes "\r". > > * hang (packet1.c:337). libssh hangs after authentication. I guess > it's because libssh doesn't run into the correct state, i.e., from > SSH_SESSION_STATE_AUTHENTICATING to SSH_SESSION_STATE_AUTHENTICATED. > Not sure if the fix is at the right place. > > * close: ssh_connection_callback (client.c:645). When in error state > (e.g., the remote server closes the connection), libssh/samplessh > doesn't close automatically. Not sure if this fix is necessary. > > * samplessh hangs (sample.c), for example, when channel_poll returns -1 (error). > > Thanks. > > - xi > > On Tue, Apr 20, 2010 at 7:32 AM, Andreas Schneider <mail@xxxxxxxxxxxx> wrote: >> On Tuesday 20 April 2010 13:05:45 you wrote: >>> Hi, >> Hi Xi, >> >>> Attached is a patch that addresses the following problems. >>> >>> * two memory leaks: enc_session (kex.c:781), session->packet_callbacks >>> (session.c:209); >>> >>> * openssl configuration problem: use OPENSSL_FOUND instead of >>> CRYPTO_FOUND, which doesn't exist in my cmake 2.8. >>> >>> * compile warnings: avoid using the name "signal" (channels.c); >>> >>> * link error when WITH_SERVER=0 (packet.c:59); >>> >>> * minor compile error (torture_options.c). >> thank you very much for your patch. I've split it up in smaller patches and >> pushed it to the repository. >> >> >> Cheers, >> >> -- andreas >> >> >>
Re: a small patch | Xi Wang <xi.wang@xxxxxxxxx> |
a small patch | Xi Wang <xi.wang@xxxxxxxxx> |
Re: a small patch | Andreas Schneider <mail@xxxxxxxxxxxx> |
Re: a small patch | Xi Wang <xi.wang@xxxxxxxxx> |